TetraTHON 2026 Winner
Sasya AI
AI-powered agriculture platform bringing voice-first advisory to smallholder farmers
Project Overview
Sasya AI is an AI-powered agriculture platform built by Team Hell Boys for TetraTHON 2026, the Indo-French international hackathon hosted by NeoFolks and Navrachana University, where it secured 1st Position among 160+ competing teams.
Smallholder farmers, who work under two hectares and make up roughly 85 percent of all farmers, rarely get advisory tuned to their own field. India loses another fifteen to twenty percent of produce after harvest to gaps in storage and market intelligence. Sasya AI closes both gaps at once with ranked seven-day crop advisories and data-driven sell, store, or transport decisions.
The platform reaches farmers and plant owners the way they already communicate, through live voice calling and a WhatsApp chatbot in local languages, backed by disease detection from a leaf photo and Mandi price intelligence.
// Tech Stack Used
Backend & APIs
Frontend
AI & Computer Vision
Conversational Access
Platform Highlights
// 01
Live Voice & WhatsApp Access
Farmers reach the platform through a live voice call or a WhatsApp chatbot in their own language, so advisory works even with low digital literacy and no app install.
// 02
Multi-Input Crop Advisory
Generates three ranked seven-day advisories for irrigation, fertilizer, and pest control using location, crop type, sowing date, live weather, and optional leaf photos.
// 03
Leaf Disease Detection
A local EfficientNet CNN classifies plant disease across 118 classes directly from a leaf image, running on our own model rather than an external vision API.
// 04
Market Intelligence
Agmarknet integration models Mandi prices, spoilage curves, and transport costs to attack the fifteen to twenty percent of produce lost after harvest.
// 05
Sell, Store or Transport
A deterministic decision engine recommends the optimal post-harvest action with an expected return projection behind every recommendation.
// 06
Grounded Generative Rationale
Gemini Flash rewrites engine output into clear, localized, farmer-facing language while prices and diagnoses stay pinned to real data, never invented.
The Impact
The whole system runs on free open data, a locally hosted vision model, and deterministic agronomy and market engines, so it scales without per-farmer inference cost. Voice and Indic-language delivery removes the literacy barrier that keeps most advisory apps unused. The first prize at an international hackathon validated the problem we chose as much as the build itself, and Sasya AI continues past the event as an ongoing product.
// Gallery
Prize Ceremony • TetraTHON 2026
Team Hell Boys receiving the winner's cheque on stage at the Indo-French AI Innovation Sprint, hosted by Navrachana University.
Winner • One Lakh Rupees
The 1,00,000 rupee winner's cheque awarded to Team Hell Boys on 7 August 2026, backed by SSIP and the Navrachana Innovation Foundation.
End-to-End Intelligent Workflow
From farmer inputs through the AI processing and model context layers to the seven-day advisory, post-harvest planner and alerts.
Crop Disease Diagnosis
A leaf photo returns the diagnosis, a confidence score and both chemical and organic treatment paths, sourced from the ICAR and KVK knowledge base.
// Team Hell Boys
- •Sumit Patel
- •Krushit Prajapati
- •Neel Prajapati
- •Patel Vrund
Hosted by NeoFolks and Navrachana University, built while studying at The Maharaja Sayajirao University of Baroda.